Advanced fluorescence based methods for protein interaction studies applied to the RNA binding protein AtGRP7

(1)

Advanced

Fluorescence based Methods

for Protein Interaction Studies

applied to the

RNA binding Protein AtGRP7

Dissertation

(2)

(3)

Inauguraldissertation

zur Erlangung der Doktorwürde

der Fakultät für Physik

der Universität Bielefeld

Advanced

Fluorescence based Methods

for Protein Interaction Studies

applied to the

RNA binding Protein AtGRP7

vorgelegt von

Fabian Humpert

geboren in Wickede

(4)

Gutachter:

Prof. Dr. Markus Sauer Prof. Dr. Thomas Huser

(5)

Zusammenfassung

Im Rahmen dieser Arbeit werden neue Methoden zur Untersuchung verschiedener Interaktionen zwischen Biomolekülen eingeführt und am Beispiel des Glyzin-reichen, RNA-bindenden Proteins AtGRP7, aus der Pflanze Arabidopsis thaliana, angewendet. Dieses Protein ist ein Forschungsschwerpunkt der Arbeitsgruppe Molekulare Zellphysiologie, mit der eine Kooperation im Rahmen des Sonderforschungsbereichs SFB 613 besteht. In transgenen Arabidopsis Pflanzen wird eine opti-mierte Variante des reversibel-schaltbaren, fluoreszierenden Proteins Dronpa (genannt Dronpa-s) mit AtGRP7 fusioniert. So können die intra-zelluläre Verteilung und die Bewegung von AtGRP7 zwischen Zell-kern und Zytoplasma, über Fluoreszenzsignale nachvollzogen werden. Unter anderem wird gezeigt, dass die Fluoreszenz gezielt im Zellkern aus und wieder an geschaltet werden kann, wodurch schließlich ein bidirektionaler Transport von AtGRP7 nachgewiesen und quantitativ charakterisiert wird.

In einem weiteren Teil der Arbeit, wird in vitro die Bindung von AtGRP7 an kurze RNA Sequenzen untersucht. Diese Eigenschaft er-möglicht es dem Protein, im Zellkern an sein Transkript zu binden, sodass die eigene Proteinbiosynthese, durch selbstregulatorische Mecha-nismen, unterdrückt wird. Hierzu werden Bindungsstudien mit Hilfe der Fluoreszenz Korrelations Spektroskopie durchgeführt. Es zeigt sich, dass AtGRP7 eine schwache aber spezifische Bindung mit einer kur-zen DNA-Sequenz eingeht, welche eine Bindestelle des Transkriptes repräsentiert.

Die RNA-bindenden Eigenschaften von AtGRP7 führen auch dazu, dass das Protein in kleinen intra-zellulären Aggregaten zu finden ist, die nur unter Stessbedingungen auftreten. Diese Proteininteraktion

(6)

kann durch Kolokalisationsstudien gezeigt werden. Herkömmliche Me-thoden zur Kolokalisationsanalyse führen hier jedoch zu nicht eindeutig interpretierbaren Ergebnissen. Als Folge daraus wird ein vollkommen neuer Ansatz zur Kolokalisationsanalyse ausgearbeitet. Dieser basiert auf der Korrelations-Matrix-Methode, die für die Detektion von Koinzi-denzen auf Einzelmolekülebene entwickelt wurde. Die Methode erweist sich als robust gegen Hintergrundsignale und ist leicht und eindeutig interpretierbar. Zusätzlich kann hiermit eine Kolokalisation von drei oder mehr unterschiedlich gekennzeichneten Objekten quantitativ ana-lysiert werden. Dies ist ein beachtliches Novum auf dem Gebiet der Kolokalisationsanalyse.

(7)

Abstract

In the framework of this dissertation, the application of different novel approaches to study protein interactions, is exemplified on the basis of the small glycine-rich RNA binding protein AtGRP7 from Arabidopsis thaliana. This protein is of special interest to the Department of Molecular Cellphysiology, which is a cooperation partner within the collaborative research centre SFB 613. An optimized variant of the reversibly switchable fluorescent protein Dronpa, termed Dronpa-s, was used as a reporter to track the protein’s localization in Arabidopsis cells, and furthermore, its intra-cellular translocation between the nucleus and the cytoplasm. Therefore, Dronpa-s was fused to AtGRP7 in trans-genic Arabidopsis plants. By switching Dronpa-s fluorescence on and off selectively inside the nucleus, a bidirectional nucleocytoplasmic shut-tling of AtGRP7 was visualized and quantified.

Within another branch of this work, the binding affinity of AtGRP7 to short RNA sequences was quantified in vitro, by means of Fluorescence Correlation Spectroscopy. Binding to RNA facilitates the protein, to interact with its transcript inside the nucleus, inducing an autoreg-ulatory inhibition to the own protein bio-synthesis. Results show a weak but specific binding of AtGRP7 to a short DNA sequence, which resembles a binding site of its own transcript.

The ability of binding to RNA helps AtGRP7 to participate in stress-related particles, termed processing bodies, upon induction of oxidative stress to Arabidopsis cells. This protein interaction was visualized by colocalization studies. Common colocalization parameters failed to reasonably interpret the here provided data. Consequently, a com-pletely novel approach to colocalization detection and quantification was pursued in this work. The here introduced analysis method defines the Γ-norm, which is based on the correlation matrix method, known

(8)

from single molecule coincidence studies, brings several benefits to the field of colocalization, like multi-color colocalization, robustness against background noise and improved interpretability.

(9)

1 Introduction

1.1 Fluorescent proteins

In today’s fluorescence microscopy the application of fluorescent pro-teins is widely used. With the discovery of GFP (Green Fluorescent Protein), Chalfie et al. [1994], Shimomura et al. [1962] and Tsien [1998] have rejuvenated research. While most organic fluorescent labels are phototoxic to living cells, fluorescent proteins can be produced intrinsically by cells as part of their protein expression cycle. Even labeling of a fluorescent reporter to a target protein, which usually proves to be difficult and error-prone, is now carried out by the cells themselves.

GFP and other fluorescent proteins (FPs) have been massively used since, to specifically tag proteins in cells. At present, the improved derivative of GFP termed eGFP (enhanced GFP) is by far the most established reporter protein in cell research. The field of application is not just limited to protein localization studies. Moreover, the mobility of proteins fused to FPs can be determined by introducing photobleach-ing to a defined region of the cell. Subsequently, the transport of other still fluorescing molecules from the surroundings into the bleached region is monitored. This method termed Fluorescence Recovery after Photo-bleaching (FRAP) was first described by Axelrod et al. [1976]. Information on protein mobility can also be gathered by a similar method which monitors the decrease of fluorescence in neighboring regions and is therefore termed fluorescence loss in photobleaching (FLIP) [Lippincott-Schwartz et al., 2001]. Since these methods are limited to the detection of secondary effects after photobleaching, a photoactivatable variant of GFP (paGFP) was engineered allowing to directly monitor the translocation of proteins fused to this reporter

(14)

[Patterson and Lippincott-Schwartz, 2002]. However, all these methods induce irreversible photo-manipulation. Recently a new generation of fluorescent proteins, which can be reversibly photoswitched between a fluorescent and a non-fluorescent state, emerged which now pave the way to novel advanced fluorescence imaging techniques. This new class of reversibly switchable fluorescent proteins (RSFPs) combines and further extends the advantages of common and photoactivatable FPs.

In the framework of the collaborative research centre SFB 613 an optimized variant of the reversible photoswitchable fluorescent pro-tein Dronpa [Ando et al., 2004] was engineered and further fused to a specific target-protein in Arabidopsis thaliana (Arabidopsis) cells. This synthesized derivative of Dronpa was termed Dronpa-s. Within the SFB 613, I therefore closely collaborated with members of the Department of Molecular Cell Physiology during my work. This in-terdisciplinary collaboration provided access to the interesting field of research on proteomics, i.e. protein structure, function and dynamics. In turn, novel fluorescence based techniques, as presented in this work, were adopted to particular and recent problems which are of special interest to the collaboration partners.

1.2 Target Protein AtGRP7

Of particular interest is the target-protein Arabidopsis thaliana glycine rich RNA-binding protein 7 (AtGRP7), which is a representative of a class of small glycine-rich RNA-binding proteins with a single RNA recognition motif (RRM). It is part of the endogenous timing sys-tem which enables the plant to pre-adopt to periodical environmental changes. Further, it participates in pathogen defense in Arabidopsis, presumably via the regulation of mRNA at the post-transcriptional level. The AtGRP7 transcript (AtGRP7 ) undergoes so-called circadian, i.e. 24-h, oscillations with a peak at the end of the daily light phase. AtGRP7 regulates these oscillations and thereby its own protein level,

(15)

1.2 Target Protein AtGRP7

by binding to the transcript and thus causing the production of an alter-native splice form with a premature termination codon. The influence of AtGRP7 on splice-site selection of its pre-mRNA obviously occurs in the nucleus where the transcript is located.

Figure 1.1: Schematic of AtGRP7 shuttling between cyto-plasm and nucleus in a plant cell.

AtGRP7 is a RNA binding protein which binds to its own

pre-mRNA. Since the pre-mRNA is located in the nucleus of the cell (N), AtGRP7 has to be imported to the nucleus from the cytoplasm (Cp). Therefore, it has to overcome the nuclear barrier.

As proteins are produced in the cytoplasm AtGRP7 has to somehow overcome the barrier surrounding the cell nucleus, for binding to its own transcript (see figure 1.1). Usually proteins which are imported into the nucleus exhibit a nuclear localization signal (NLS), a special amino acid sequence. The NLS enables strong binding to importin, a protein which helps other proteins to cross the nuclear barrier. Albeit AtGRP7 lacks a classical NLS, it exhibits a sequence simmilar to a different known nuclear import signal, an M9 domain originally found in mammalian heterogeneous nuclear ribonucleoproteins (hnRNPs). In mammalian cells M9 domains have been reported to facilitate both transport into and out of the nucleus [Bogerd et al., 1999].

The molecular mechanisms of the nuclear import have not been re-vealed and at present it is not known whether AtGRP7 is also

(16)

ex-ported from the nucleus to the cytoplasm. This bears the obvious question, if AtGRP7 can get into the nucleus just by diffusion or by active transport processes, involving an M9 domain equivalent in plants.

1.3 Application of photoswitchable proteins

in standard imaging techniques

To analyze such protein interactions, in cell biology often fluorescence imaging techniques like the aforementioned FRAP are used. The general problem to this kind of analysis is, that cells are exposed to high laser intensities to efficiently photo-bleach the FPs, often disturbing their physiological function. Photoswitching of RSFPs to the off-state requires much less photons than photobleaching eGFP, allowing FRAP or photoactivation measurements to be performed at much lower excitation intensities, and therefore, inducing less phototoxicity to the cell [Ando et al., 2004]. This work aims to answer the question if RSFPs, like Dronpa-s, can easily be introduced to standard imaging techniques, to replace commonly used FPs efficiently.

1.4 Colocalization analysis visualizing

protein interactions

Colocalization studies represent another approach to unravel protein interactions by just providing information on spatial proximity of inter-action partners. Therefore, different fluorescent labels are attached to the interaction partners of interest. Thereby, avoiding spectral overlap and any other interactions between the used fluorophores is demanded. Typically, fluorescence images of the such prepared sample are recorded on spectrally separated channels, preferably at the same time. Now, by overlaying the images of the different channels, a merged image denotes sites of colocalization by an additive secondary color. That is, two

(17)

1.4 Colocalization analysis visualizing protein interactions

merged channels of the primary colors red and green yield the secondary color yellow as a reporter for colocalization.

Figure 1.2: Example of the distraction of human vision by colors an shapes.

Most people with healthy vision are consumed by the illusion that the prominent spirals are in shades of green and blue, but they actually have the very same color. [Sample image provided by Fiji]

Due to the shortcomings of human vision, judging by shapes and colors does not suffice to provide reliable information on colocalization. Figure 1.2 demonstrates the problems to be encountered. Approaches to provide an unbiased numerical parameter to indicate and quantify colocalization led to the institution of the at present mainly used Pearson Correlation Coefficient (PCC) and Manders Overlap Coef-ficient (MOC)[Manders et al., 1993]. Both are not straightforward to apply correctly to image data, and thus, often lead to ambiguous results due to misinterpretation or vague definition of colocalization and noncolocalization. Up to now all implementations of colocalization approaches are limited to determine colocalization between a maximum of two channels. However, in many cases more than two interaction partners are contributing to a specific peculiarity. Thus they have to be examined crosswise. The work at hand presents advances in overcoming these limitations.

(18)

(19)

2 Theoretical Background

Due to the interdisciplinary nature of this work, the theoretical back-grounds of both, biological and physical subtopics, have to be pointed out. In the following sections a short introduction to fluorescence itself and fluorescence based research methods will be presented. Fur-thermore the main aspects of the examined biological systems will be mapped in chapter 2.4.

2.1 Principles of Fluorescence

Luminescence being the superordinate effect (besides black body radi-ation) of light emission by molecules, can be subdivided into several subtypes, i.e. radioluminescence, chemiluminescence and photolu-minescence, depending on which effect causes the emission of pho-tons.

Photoluminescence can occur, when a molecule absorbs a photon of the energy

E = h · ν (2.1)

with the frequency ν and the Planck constant h. If E is equal or greater than the energy difference ∆E between the ground state S0 and the first electronically excited state S1 of the molecule.

The light absorption occurs on a 10-15 _{seconds time scale. Relative to} the nuclear movement (10-13 _{seconds) these electronic transitions occur} almost instantly which is the keynote to the Born-Oppenheimer

(20)

approximation and also accounts for the vertical transitions in the Franck-Condon energy diagramm (figure 2.2).

The excited state endures some 10-9 _{seconds until the molecule relaxes} to the ground state by emitting a photon. The fluorescence lifetime is specific for different fluorophores.

The emitted fluorescence photon carries the energy E∗_{, with}

E − E− = E∗. (2.2)

Therefore, the maintained energy is composed of

E∗ = h · ν∗, (2.3)

and

ν∗ ≤ ν. (2.4)

applies.

Depending on the energy of the incident photon and the energy gaps, the molecule is most often not only lifted to the lowest possible energy state, but to a higher excited vibrational state of the 1st _(S

1) or nth (Sn) excited electronic state.

However, according to Kasha’s rule fluorescence emission occurs from the lowest electronically excited state, because the energy gaps separating higher excitation levels are close enough for the molecule to decay by internal conversion, thus increasing the thermal energy of the surrounding solution. This also explains the common experience of the emission wavelength being independent from the excitation wavelength.

While this rule has some exceptions, it applies for the most molecules. Therefore, the emitted light is commonly red-shifted to the incident light, due to the thermal relaxation prior to the fluorescence process, which is called the Stokes Shift.

(21)

2.1 Principles of Fluorescence

Figure 2.1 shows a transition from an excited electronic singlet state to an energetically lower triplet state, termed intersystem crossing. This transition implies the need of a change in spin orientation, which is quantum mechanically forbidden.

While the electron is in the triplet state, it can relax to the ground state by undergoing another spin transition and emission of radiation, termed phosphorescence. Since this, again, is not allowed, the lifetime of this state is long compared to fluorescence.

Apart from intersystem crossing, there are other concurring processes, which prevent the emission of fluorescence. These processes are termed

S

₀

S

₁

S

n

T

₁

E

ne

rgy

Intersystem Crossin g Fl uo resce nce A bso rptio n P ho sph or escen ce Inter na l Co nve rsio n

Figure 2.1: Jablonski diagram.

The diagram illustrates the electronic states of a fluorophore and the transitions between them. The vibrational ground states of each electronic state are indicated by thick horizontal lines, the higher vibrational states by thin lines, whereas the electron spin orientation is given by small arrows. Nonradiative relaxations are displayed as black or gray arrows between electronic states. Along with all nonradiative relaxations the phosphorescence, which is the radiative relaxation from the triplet state (T₁) to the singlet ground state (S0), is competitive to the fluorescence relaxation.

(22)

quenching mechanisms, which will be described in chapter 2.2.2 on page 15. All the concurring processes can be described as the rate kN R of nonradiative relaxation processes from S1 to S0. In analogy the rate of radiative relaxation processes can be defined as kR, being the fluorescence rate. The sum of these rates gives the inverse of the average time a molecule spends in the first excited state, also termed fluorescence lifetime τF l.

τF l= 1 kR+ kN R

(2.5) For the absorption and emission of photons, the Franck-Condon principle gives information on probabilities for the respective transi-tions, plotted vertically in the Franck-Condon diagram, depicted in figure 2.2. The transitions will occur most likely, if the two vibrational

E

ne

rgy

r

Figure 2.2: Franck-Condon diagram.

Possible transitions between electronic states of a molecule most often occur between overlaps of vibrational wave functions. Tran-sitions happen too fast for the nuclear coordinates r to change.

(23)

2.2 Fluorescence Spectroscopy

of absorption- and emission spectra known from most fluorescing com-pounds. The varying transition probabilities result in an inhomoge-neous distribution of photon energies, that are absorbed and emitted, as can be seen in figure 2.3.

Intensity

Energy

0 6

0 5

0 4

0 2

0 3

0 1

0 0

0 6

5 0

4 0

2 0

3 0

1 0

6 0

Fluorescence

Absorption

Figure 2.3: Probabilities of electronic transitions in ab-sorption and emission processes (cf. figure 2.2).

Exci-tation and emission probabilities are mirrored, as well as the resulting spectra.

2.2 Fluorescence Spectroscopy

Fluorescence spectroscopy aims to identify and/or quantify molecules by their fluorescence characteristics. This is rendered possible, as each molecule has characteristic absorption and emission spectra due to the specific composition of electronic and vibrational states.

A class of molecules which feature most suitable fluorescence char-acteristics for spectroscopic purposes is highlighted in the following section.

(24)

2.2.1 Fluorophores

Molecules considered as fluorophores meet special needs for efficient flu-orescence generation an detection. Therefore they feature a reasonable quantum yield Φ, defined as ratio of emitted fluorescence photons to absorbed excitation photons. Φ can also be defined with the relaxation rates given in equation (2.5):

Φ = kR kR− kN R

(2.6) Depending on the method of detection, the spectral range of fluores-cence can be of importance. Most fluorophores emit visible or near visible light. 500 600 700 800 Absor ption Wavelengthf[nm] Fluor escence Stokes-shift

Figure 2.4: Absorption and emission spectra of the

artificial fluorophore Atto655 [AttoTec].

Besides the many natural organic dyes, synthetic fluorophores have been designed. Therefore, a wide range of the spectrum is crowded by according fluorescence emitters. Especially the visible spectrum is consistently covered. Figure 2.4 shows absorption and emission spectra of one of these fluorophores, named Atto 655 (AttoTec, Germany). The Stokes shift (see chapter 2.1) of 19 nm between absorption and emission maxima, can clearly be seen.

(25)

While the most molecules of interest to natural sciences have inferior fluorescence abilities, fluorophores are attached to those molecules as a fluorescent marker. Thus identification and quantification of even non-fluorescing molecules can be afforded by fluorescence spec-troscopy.

2.2.1.1 Fluorescent Proteins

Attaching a fluorescent marker to a target molecule can be a diffi-cult task under certain conditions. Most problems evolve from the urge to mark bio-molecules inside living cells, sustaining cell life at best. Therefore it was a revolution to natural science, when Osamu Shimomura was to discover the Green Fluorescent Protein (GFP) in 1962. Being a protein from the jellyfish Aequorea victoria, a gene in the Aequorea genome encodes for GFP. It was just a question of time, until the gene’s nucleotide sequence was published by Prasher et al. [1992]. Two years later Chalfie et al. [1994] succeeded in ex-pressing recombinant GFP in E. coli and C. elegans. Eventually the (by then poor) spectroscopic abilities of GFP were enhanced in the lab of Roger Tsien [Heim et al., 1995]. From there GFP raised its importance as a widely-used tool in natural sciences. Several GFP mutants were engineered to feature either further enhancement or even different emission wavelengths. Thus, by the beginning of the 21st century a broad spectrum of fluorescing proteins, emitting various colors of light, were available.

All these fluorescent proteins have the huge advantage of not only being applicable for in vivo applications but furthermore, enabling intrinsic tagging bio-molecules with a fluorescent reporter by the organism itself. Therefore, the fluorescent reporter is genetically fused to a target protein inside a transgenic organism. This fusion protein is translated in one go.

Besides all advantages, fluorescing proteins have general drawbacks compared to synthetic fluorophores, like low quantum yields and poor photostability. Also they are limited in size. While not all

(26)

of them are available as monomers, even a monomeric GFP-like pro-tein is of reasonable size. At the margin this can result in fusion proteins, where the reporter is several times bigger than the target. This may alter the target’s properties immensely, yielding artifact results.

Figure 2.5: Stereoview of a Dronpa molecule.

Dronpa molecule (PDB-2IOV) was visualized with PyMOL [Andresen et al., 2007, DeLano, 2002]. The common barrel structure is shared by most fluorescing proteins. Here the ’barrel’ consists of 11 β-sheets enclosing the chromophore tripeptide which is pictured in detail in figure 2.6. 3D immersion by crosseye stereoview.

In 2004 Ando et. al. could deliver a photoswitchable fluorescent protein, which was named Dronpa after dron, a ninja term for vanishing, and pa for photo-activation. This protein can be reversibly switched to a dim state featuring low fluorescence and a bright state with high fluorescence intensities. The switching occurs by irradiation with light of different wavelength as shown in figure 2.6.

(27)

Figure 2.6: Stick model showing light-induced cis/trans iso-merization of the chromophore in Dronpa.

The chromophore of Dronpa is formed by the tripeptide Cys-62–Tyr-63–Gly-64 (CYG). In the fluorescent state the chromophore adopts a

cis-conformation, while it adopts the trans-conformation in the dim

state. The isomerization is induced by light of different wavelengths. Thus blue light induces transition to the dim state and UV light again restores the bright state.

2.2.2 Fluorescence Quenching

As described before, processes which diminish fluorescence are called quenching. If another molecule is involved by causing the quenching, this molecule is referred to as quencher. Collisional or dynamic quench-ing, occurs when an fluorophore, while excited, hits another molecule in the solution. Often molecular oxygen acts as quencher this way [Kautsky, 1939].

The process of dynamic quenching can be described by the Stern-Volmer equation:

F0

F = 1 + K · [Q] = 1 + kq· τ0·[Q], (2.7) where F0 and F are the fluorescence intensities with and without quencher respectively. K is the Stern-Volmer quenching constant, [Q]

(28)

the concentration of the quencher molecules, kq the quencher rate coeffi-cient and τ0the lifetime of the un-quenched fluorophore.

Furthermore, fluorophores and quenchers can build non-fluorescent complexes, which is termed static quenching. This can be described by

F0

F = 1 + KS·[Q], (2.8)

with the association constant KS. Which can be expressed by KS = [F

0] [F ] · [Q] −

1

[Q]. (2.9)

Photoinduced electron transfer (PET) is one of various quenching processes.

In PET redox interactions between a fluorophore and its quencher yield a two stepped electron transfer. This demands a very close distance between both molecules [Marcus and Sutin, 1985]. This is emphasized by the Rehm-Weller equation

∆GCS = Eox− Ered− E0,0−

e2

ε · d (2.10)

where ∆GCS is the discharged energy, Eox the oxidation potential, Ered the reduction potential and E0,0 the energy of the electron transition from S0 to S1 state. The term e

2

ε·d is the solvent effect with ε the dielec-tric constant and d the charge separation distance.

By excitation the former lowest unoccupied molecular orbital (LUMO) of the fluorophore, it turns into a singly occupied molecular orbital (SOMO). Next the fluorophore is reduced or oxidized by the quencher.

In case of oxidation the electron on the SOMO of the fluorophore is transferred to the LUMO of the quencher (1), thus turning it into a SOMO, too. From the SOMO of the quencher the electron can be transferred to the former highest occupied molecular orbital (HOMO)

(29)

2.2 Fluorescence Spectroscopy LUMO LUMO HOMO HOMO

Reduction

LUMO LUMO HOMO HOMO

Oxidation

hν

1 2 2 1

Ener

gy

_hν

Figure 2.7: Photoinduced electron transfer between fluorophore and quencher. By excitation an electron on the highest occupied molecular

orbital (HOMO) of the fluorophore is promoted to the lowest unoccupied molecular orbital (LUMO). After that the fluorophore can be oxidized or reduced by an appropriate quencher in close contact. Two non-radiative electron transitions occur, restoring the initial state of both molecules thereby.

of the fluorophore (2), thereby restoring the initial state. In case of reduction an electron is promoted from the HOMO of the quencher to the former HOMO of the fluorophore (1), as depicted in figure 2.7. Now the electron on the SOMO of the fluorophore restores the initial state by migration to the former HOMO of the quencher (2). All of these electron transitions are radiation-less processes.

2.2.3 Fluorescence Correlation Spectroscopy

The advantage of kinetics on the molecular level, arising on character-istic time scales, is taken by Fluorescence Correlation Spectroscopy (FCS) [Magde et al., 1974]. In figure 2.8 those kinetics are depicted on the respective time scales. Beginning with antibunching, the rate at which fluorescence photons can be emitted by the fluorophore, and solvent reactions or conformational changes, on the sub-ns to ns time scale. Furthermore, photophysical processes, like triplet transition,

(30)

ms µs ns Conformational Kinetics Photophysics Diffusion Antibunching

G(τ)

log(τ)

Figure 2.8: Autocorrelation function used for FCS analysis.

The sketched function spans multiple time scales from the nano second regime to seconds. Different kinetics can be identified on the respective time scales.

mainly occur in the µs regime, whereas diffusional processes take place in the ms range.

The most common approach to FCS measurements, can be realized on a confocal setup, as described in chapter 2.3 on the next page. Assuming a Gaussian intensity profile, the excitation volume can be approximated by I(~r) = I0exp − 2(x2+ y2) (ωxy)2 − 2z 2 (ωz)2 ! , (2.11)

with the axial and lateral diameters of the excitation volume ωxy and ωz respectively.

(31)

2.3 Fluorescence Microscopy G(τ) = ¯N−1· 1 + τ τD −1 1 + τ ω2_τ D −1 2 Yn i=1 1 + Ki·exp τ τKi , (2.12)

where ¯N is the mean number of fluorescing particles in the detection volume, τD the characteristic diffusion time and the aspect ratio of the excitation volume ω = ωz

ωxy. Ki = k_AiBi

k_BiAi are the equilibrium constants and τKi = (kAiBi + kBiAi)

−1 _{the characteristic times for n additional} processes of interest (cf. figure 2.8 with n = 3).

The signal to noise ratio (SNR) of a FCS experiment is determined by

SNR = q G(τ)

var (G(τ)) ≈ G(τ)ν ¯N ∝ ν √

T , (2.13)

with the mean number of photons emitted by each molecule per time interval ν, and T the total number of time intervals during the experi-ment. As can be seen, the SNR improves with the total duration of the experiment and the number of photons per particle. The latter is highly dependent on the quantum yield of the molecule or its fluorescent marker (see chapter 2.2.1 on page 12).

2.3 Fluorescence Microscopy

The unaided human eye can barely resolve structures of less than half a millimeter. For resolving smaller objects, some kind of microscope has to be applied. But also light microscopy has its limits in resolution. This manifests in the diffraction limit, found by Ernst Abbe in 1873 as

d= λ 2NA =

λ

(32)

with the radius of a minimal projected spot d, the wavelength λ and the numerical aperture of the objective NA. Where α is the opening angel and n the refractive index of the optical medium separating objective and sample.

This also complies with the Rayleigh criterion. The smallest possible point light source displayed by a circular aperture yields a projection in form of an Airy disc. The lateral intensity profile is then termed point spread function (PSF), which is mathematically described by a Bessel function of first kind and first order. The distance between the main maximum and the first minimum of a PSF is the Rayleigh criterion dAiry = 0.61λ NA ≈ λ 2NA, (2.15)

which defines the smallest possible, still resolvable distance between two point light sources (see figure 2.9).

x y x z y I dAiry dz

Figure 2.9: Diffraction image of a point light source.

The x-y plane exhibits the charac-teristic Airy disc pattern. Intensity profiles are given on the right, with respective distances denoted, which define the lateral and axial resolu-tion.

With the three-dimensional diffraction image of a point light source the axial resolution can be defined analogously, again by the axial distance from the central maximum to the first minimum

(33)

2.3 Fluorescence Microscopy

dz = λ ·2n

NA2 (2.16)

Due to that in microscopy two parameters are pushed to the limits for enhancing resolution. First of which, the NA of an objective, has reached typical values of 1.4 in research and even 1.7 NA objectives are available by now. The second parameter λ was pushed to new limits by changing from light microscopy to electron microscopy, as electrons have a by four magnitudes smaller wavelength than pho-tons.

However, in live cell imaging, fluorescence microscopy still is state of the art and commonly used, as shorter wavelengths are always a trade-off between higher resolution and harming the living system by high energy radiation. Hence, there is a lot of effort to further increase the resolution in fluorescence microscopy.

This can be achieved by the limitation of the field of view or detection volume. One approach to realize a thus enhanced resolution is the confocal laser scanning microscopy (CLSM).

The Confocal Principle

In CLSM a laser beam is focused on the sample by an objective. The emitted fluorescence is collected by the same objective and focused on the aperture of a pinhole. Thereafter the fluorescence light is collimated and focused on a single point detector. By the pinhole a small detection volume is defined in overlapping with the focal volume. Thus, only fluorescence and scattered light out from the detection volume can pass the pinhole to reach the detector, as pictured in figure 2.10.

For imaging purposes the sample has to be scanned. That way an image can be constructed out of the intensity values for each scanned coordinate in three dimensions. Therefore, 2D or 3D images of the sample can be generated. The scanning is done either by moving the laser beam over a static sample, or moving the sample over a static beam.

(34)

Figure 2.10: Standard confocal microscope.

Laser light is focused by an objective (Obj) on a sample. Fluorescence light is collected by the same objective. A dichroic mirror (DCM) separates the laser light from the red-shifted fluorescence light (cf. Stokes shift in chapter 2.1). A lens (L) focuses the fluorescence light on a pinhole, which acts as a spacial filter. Finally, it is focused again on a point-detector.

The resolution in confocal microscopy is enhanced by a factor of 1.41 in relation to standard wide field microscopy.

Another approach to limit the detection volume is total internal reflec-tion (TIRF) microscopy .

Total Internal Reflection Fluorescence Microscopy

If light passes through two different optical media, it changes its direction at the interface, according to the difference in the refrac-tive indices of the two media. This is described by the Snell’s law: n1 n2 = sin(α2) sin(α1) , (2.17)

with ni the refractive indices of the media and αi the beam angles in the different media, as sketched in figure 2.11.

(35)

2.3 Fluorescence Microscopy α1 α2 n₂ n1 Figure 2.11: Refraction of light upon change of refrac-tive indices.

The largest possible angle of incidence which still yields a refracted ray of light is referred to as the critical angle. Under this condition the refracted ray propagates along the interface of the two optical media. The incident light gets totally reflected, as the angle is further increased. In this case an evanescent wave in the medium with lower refractive index emerges. The penetration depth of this wave is given by d= λ0 4π · (n2)2· sin(α2) − (n1)2 −1₂ , (2.18)

with the vacuum wavelength λ0 of incident light. The intensity of the evanescent wave is a function of the distance z from the inter-face

I(z) = I0· e

−z

d , (2.19)

where I0 is the initial intensity at the interface .

In TIRF microscopy this is utilized to limit the excitation volume. Therefore, a laser beam is focused on the back focal plane of an objective with a high NA. When the beam is then shifted close to the margin of the objectives aperture, it is deflected and collimated. Eventually the beam is totally reflected at the interface, under an incident angle bigger than the critical angle. This results in an evanescent field, emerging in

(36)

the sample, with a penetration depth of only some 100 nm, depending on the parameters mentioned above.

(37)

2.4 Biological Background

A quite remarkable part of this work findings in molecular biology are presented. Thus a short summary of the corresponding fundamentals is given in chapter 2.4.1 and chapter 2.4.2. First of which gives an introduction to the structure and function of nucleic acids, whereas the second delivers insight into the synthesis of proteins within a cell.

2.4.1 Structure of Nucleic Acids

O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O N N N N N N N N N N N N _N N N N N N N N O_ O_ O_ O_ O_ _O _O _O _O _O P P P P P P P P NH2 OH OH NH H2N HN NH2 H2N HN H2N NH NH2 5' 5' 3' 3' A T G C A T G C O O O O O O O O O O O O O O O O O O O N N N N N N N N N N O_ _O _O _O _O P P P P NH2 OH NH NH2 NH NH2 5' 3' A C U G OH dsDNA ssRNA

Figure 2.12: Chemical Structure of DNA and RNA.

Bases are identified by the encircled abbreviations. In RNA the thymine (T) is replaced by uracil (U). H-bonds are denoted by dotted lines.

(38)

Desoxyribonucleic acid (DNA) and ribonucleic acid (RNA) are to be found in every single organism, providing the storage and delivery of genetic information. Both types of nucleic acids consist of nucleotide se-quences and are therefore called polynucleotides1.

Each nucleotide is composed of a backbone, consisting of a sugar2 and a phosphate group, and a nucleobase attached to the sugar. This structure gives rise to the pairing of two complementary nucleobases, as hydrogen bonds (h-bonds) form amongst them.

In DNA the sequence of the nucleobases adenine (A), cytosine (C), guanine (G) and thymine (T) encodes genetic information. In RNA thymine is replaced by uracil (U). The nucleobases are subdivided in pyrimidines (Y) and purines (R), while A and G are purines and C, T and U pyrimidines. Thus the complement to a pyrimidine is a purine, which leads to the pairing of A and T3 _{with two h-bonds as} well as G and C with three h-bonds in between (cf. figure 2.12). This most common binding pattern is termed Watson-Crick base pairing. However, other binding patterns exist4_{, yielding more complex tertiary} structures.

In most cases the DNA is available as a double strand (ds) composed of two anti-parallel DNA strands bound to each other due to base pairing of complementary nucleobases. Each DNA strand has two ends which differ in orientation of the terminal sugar molecule. While one end exhibits the 3’ carbon the other end exhibits the 5’ carbon, as depicted in figure 2.12. Thus each polynucleotide strand features one 3’-end and one 5’-end. Hereby the direction of a strand can be defined. In dsDNA the two complementary strands are of opposite direction. This gives rise to the double-helical structure of DNA (see figure 2.13 a) which was first discovered by James Watson and Francis Crick in 1953.

In contrary to DNA RNA is most often available as a single strand (ss). 1_{Short strands of nucleotides are called oligonucleotides (oligos) accordingly.} 2_{2-desoxyribose in DNA and ribose in RNA}

3_{U in RNA respectively} 4

(39)

H-bonds between base pairs in ssRNA account to the building of sec-ondary structures like hairpin-loops (see figure 2.13 b).

Figure 2.13: Stereoview of DNA double helix and RNA hairpin-loop.

DNA and RNA struc-tures were calculated with make-na server and MC-fold respectively. The generated models were visualized with PyMOL [DeLano, 2002, Parisien and Major, 2008]. 3D immersion by crosseye stereoview.

a

b

2.4.2 Protein Biosynthesis

Protein biosynthesis is a multi-step process. First the genetic in-formation is transcribed from DNA to RNA which, in a second step, is translated into aminoacid sequences. These are the building blocks of proteins. The whole process is also termed gene expres-sion.

(40)

2.4.2.1 Transcription

The transcription is performed by the RNA polymerase (RNAP)5_, which binds to one of both strands in dsDNA, the so called template strand6. This binding region is called promoter. Transcription starts from the 3’-end of the promoter, thus the opposite coding strand7 of the DNA is copied. As a result the pre-mRNA is assembled from 5’-end to 3’-end.

If a gene codes for a protein the pre-mRNA is termed transcript of the protein. Genes can be enhanced or silenced, meaning the expres-sion of the gene is promoted or suppressed respectively. Therefore, regulatory sequences termed untranslated regions (UTR) upstream (5’UTR) and downstream (3’UTR) of the coding region in a DNA strand can direct the gene expression accordingly. These regions are also transcribed to the pre-mRNA and, as they lack gene coding information.

Splicing

In eukaryotic cells a further step in RNA processing can occur. If the pre-mRNA sequence between 5’UTR and 3’UTR contains introns a process termed splicing takes place. In the process of splicing the intron regions are removed, leaving the remaining exon regions to build up the thereby matured mRNA, which is illustrated in figure 2.14. The mature mRNA can be translated to a protein in a following process. In eukaryotic cells the complete transcription takes place inside the nucleus, the organelle which holds the DNA.

For the final protein synthesis the mature mRNA has to leave the nucleus and enter the cytoplasm of the cell, where the translation takes place.

5_{In eukaryotic cells there is more than just a single type of RNAP.} 6_{Also termed anti-sense strand.}

(41)

Intron

5' UTR

Exon

3' UTR

Exon

5' UTR

Exon

pre-mRNA

mature mRNA

Figure 2.14: Maturation of pre-mRNA by splicing.

Depicted is a pre-mRNA strand with a 5’ exon, an intermediate intron and a 3’ exon. The non-coding intron is cut out, while the remaining exons build up the mature mRNA.

2.4.2.2 Translation

As polypeptides or proteins consist of amino acids, the nucleic al-phabet of four characters has to be translated into the amino acid alphabet. While there are 22 proteinogenic amino acids which can build proteins, only 20 of them are directly encoded. Thus, the amino acid alphabet consists of 20 characters. A translation from genetic code into amino acid code is warranted, hence a specific sequence of three nucleotides ,termed codon, each encodes for a specific amino acid yielding 43 _{= 64 different possible codons. Only three of all possible} codons do not encode for amino acids. Therefore, most amino acids are encoded by more than one codon. Thus, the genetic code is referred to as degenerated. However, most organisms feature a characteris-tic preference of one or two codons over the other codons to encode for a specific amino acid, which is reflected in the organism’s DNA sequence.

The three codons not encoding for an amino acid, are termed stop-codons8, as they induce the termination of the polypeptide chain elongation in the C-terminal end featuring a free carboxyl group. In analogy the other end of the polypeptide is termed N-terminal end, as it exhibits a free amine group.

(42)

2.4.3 Circadian Rhythms in Plants

Virtually all organisms have to adapt to the 24 hour light-dark cycle in consequence to the rotation of the earth. Consequently many higher organisms evolved internal clocks with circadian rhythms which enable them to pre-adapt to periodic environmental changes. On exemplary benefit plants are enabled to synthesize all proteins needed for photosynthesis right before dawning, thus prolonging the period of effective photosynthesis during daylight.

The complex structure of the internal clock can be abstracted by defin-ing three greater sections. These sections termed input pathway, main oscillatorand output-pathway are sketched in figure 2.15.

Figure 2.15: Scheme of the internal clock of Arabidopsis.

The internal clock is subdivided into an input pathway, where external stimuli start resp. reset the clock, an oscillator, setting the pace, and the output pathway, where the phenotypic response is generated.

Here the input pathway describes the reception of endogenous signals mainly in form of light and temperature. These signals are carried further downstream to the main oscillator, where the signal may induce the expression of so called clock proteins. These proteins interact with each other, generating a circadian rhythm of about 24 hours. This oscillation is directed downstream, promoting expression of clock related proteins like AtGRP7, which possibly generate oscillation

(43)

themselves, thus being termed slave-oscillators. All these oscillating signals eventually trigger phenotypic response, preparing the plant for changes in the environmental conditions.

On the molecular level the oscillations of the main oscillator are gener-ally generated by feedback-loops. These are afforded by enabling or disabling gene expression on the transcriptional level, regulation on the posttranscriptional level by i.e. alternative splicing, or direct protein modification on the posttranslational level.

In Arabidopsis, a largely used model organism for higher plants in general, the master-oscillator consists of the main clock proteins LHY (Late Elongated Hypocotyl), CCA1 (Circadian Clock Associated 1) and TOC1 (Time of CAB Expression 1). These proteins interact in interlaced positive and negative feedback loops. TOC1 exhibits a maximum expression level in the evening hours, due to the so called evening event (EE) embedded inside its promoter (cf. chapter 2.4.2.1 on page 28). The EE is built by the sequence AAATAACT which can be found in the majority of genes with their expression peak in the evening [Harmer et al., 2001]. The expression of both LHY and CCA1 is induced by TOC1, resulting in a maximum expression during the morning hours, and therefore, phase shifted to that of TOC1. Furthermore, CCA1 and LHY both bind to the TOC1 promoter, causing a repression of TOC1 protein expression [Alabadí et al., 2002]. However, LHY and CCA1 additionally repress their own as well as the other’s gene expression, again by binding to the respective promoter. Besides this feedback-loop there are further affiliated regulation loops called morning loop and evening loop [see Farré et al., 2005, Fujiwara et al., 2008, Kim et al., 2007, Locke et al., 2005, Para et al., 2007, for futher reading]. These multiple regulation pathways yield the stable contrariwise oscillation of the three proteins throughout the day.

The next section introduces AtGRP7, a small protein whose transcrip-tion is subject to this circadian rhythmicity.

(44)

2.4.4 Glycine-rich RNA-binding Proteins in

Arabidopsis

AtGRP7 is a small glycine-rich RNA binding protein in Arabidop-sis. RNA binding proteins are highly conserved in plants, animals and also in the human genome. They regulate gene expression on the post transcriptional level and influence maturing, modification and degradation of mRNA. Schmidt et al. [2010] described this reg-ulation processes to aid the plant in resisting abiotic and biotic stress.

Under cold conditions AtGRP7’s gene expression is induced. The protein increases stress tolerance by causing the closing of stomata.

Figure 2.16: Predicted structure of AtGRP7 without its glycine stretch.

The protein’s binding pocket is built up of four β-sheets (red) which are lined up by two α-helices (yellow). Connecting random coil amino sequences are also shown (blue). The two centric β-sheets hold a RRM sequence each (cf. Sequence 2). The structure was predicted by homol-ogy modeling with CPHmodels [Nielsen et al., 2010]. 3D immersion by crosseye stereoview.

(45)

However, under dehydration and salt stress conditions AtGRP7 pro-motes the opening of stomata.

Furthermore, a role in immune response mediation is assumed, because loss of functionmutants show an increased susceptibility to pathogens. Along with that, a reduction of mRNA export from the nucleus was observed, as well as late flowering of the mutants, suggesting an inter-action with the autonomous flowering pathway.

AtGRP7’s function in the immune response pathway is presumably aided by the influence of the circadian rhythmicity, as it was found to be a clock related protein itself and part of a so called slave oscillator down-stream the master oscillator [see Heintzen et al., 1997].

Staiger and Apel [1999] found the promoter of AtGRP7 to feature two domains giving rise to the rhythmic expression of its protein. Later on one of these domains was identified as EE by Harmer et al. [2001].

AtGRP7 regulates the circadian oscillations of its own transcript At-GRP7. The transcript concentration in Arabidopsis peaks 8-12 hours after illumination (input pathway), while AtGRP7 protein achieves peak levels 4 hours later. This shifted oscillation indicates a negative auto-regulation that is considered to be initiated by AtGRP7 binding to its own pre-mRNA (cf. chapter 2.4.2.1 on page 28). This binding induces an alternative splicing, which results in an alternate transcript featuring a premature stopcodon (cf. chapter 2.4.2.2 on page 29). Thus the translation is aborted too early, yielding a non-functional protein, which is sketched in figure 2.17.

Furthermore, the transcript is subject to the nonsense mediated decay showing a shorter half life than the correctly spliced transcript, as described by Schöning et al. [2008]. Staiger et al. [2003] identified this negative auto-regulation as the propulsive force according for the per-sisting oscillation of AtGRP7 and its transcript.

The correctly spliced and translated AtGRP7 features a N-terminal RRM and a glycine stretch at the C-terminus. The RRM is built up by the two Ribonucleoprotein (RNP) domains RNP1 and RNP2, and ac-counts for binding to the pre-mRNA. Furthermore, the glycine stretch

(46)

Intron 5' 3' pre-mRNA 5' 3' Splicing mature mRNA Translation functional protein AtGRP7 Intron 5' 3' AtGRP7 bound to pre-mRNA 5' STOP 3' Splicing alternatively spliced mRNA Translation truncate protein

Figure 2.17: Alternative Splicing of AtGRP7 in comparison to correct splicing.

AtGRP7 induces alternative splicing by binding to its own transcript (top

right), resulting in a mRNA with a premature stopcodon, which is translated to a non-functional truncated protein. Thus AtGRP7 negatively regulates its own expression.

holds a M9 domain, which accounts for intra-cellular transport of AtGRP7, as described by Ziemienowicz et al. [2003].

Figure 2.16 shows the predicted structure of AtGRP7 omitting the glycine stretch, as no sufficient data exists for structural elucidation. RNP1 and RNP2 each locate to one of the two central β-sheets, forming the binding motif.

The amino acid sequence of AtGRP7 is shown in Sequence 2, with highlighted RNPs and the predicted domain. Sequence 1 shows the nucleotide sequence of AtGRP7 with highlighted binding sites accord-ingly. Prior to this work, two binding sites have been roughly identified via depletion analysis and Electrophoretic mobility shift assay (EMSA) [Schöning et al., 2007, Staiger et al., 2003]. They locate to the 3’UTR and the second half of the intron. The 3’UTR binding site was further investigated by Schüttpelz et al. [2008].

(47)

Sequence 1: AtGRP7 nucleotide sequence.

UTR is displayed in gray, binding regions in orange and exons in blue. 1 CUUCGUCUAC AUCGUUCUAC ACAUCUCACU GCUCACUACU CUCACUGUAA 51 UCCCUUAGAU CUUCUUUUCA AAUUUCAAUG GCGUCCGGUG AUGUUGAGUA 101 UCGGUGCUUC GUUGGAGGUC UAGCAUGGGC CACUGAUGAC AGAGCUCUUG 151 AGACUGCCUU CGCUCAAUAC GGCGACGUUA UUGAUUCCAA GGUCUGUUAC 201 ACGCCGAGAU CGGACUCCGA GUGAUAUCGA UGAUCUCAUC CUCGACGGAU 251 CUGUUCCGAU CUUGUGUUUC UCUGUUACUU GAUUCGAUUA CUCUGUUACU 301 AUUCUCGUUC UUUGUUACUA CUACUACUAC UACUGUUACU UGUAUUUUCC 351 CAAAUCGGUA CGUUCAUCUU CCUGCUUCUG UGAGCCCGGA GAUCGAUCGG 401 AUUUUUUUGU AUUUUGUAUA UUUGUUGUAG AUCUAAAUGC UUUUGUUCAG 451 UUUUGUUGGA UUGUUUUGCU GAUCUGGUUU UUGUAUUAUU UGGAUAACAG 501 AUCAUUAACG AUCGUGAGAC UGGAAGAUCA AGGGGAUUCG GAUUCGUCAC 551 CUUCAAGGAU GAGAAAGCCA UGAAGGAUGC GAUUGAGGGA AUGAACGGAC 601 AAGAUCUCGA UGGCCGUAGC AUCACUGUUA ACGAGGCUCA GUCACGAGGA 651 AGCGGUGGCG GCGGAGGCCA CCGUGGAGGU GGUGGCGGUG GAUACCGCAG 701 CGGCGGUGGU GGAGGUUACU CCGGUGGAGG UGGUAGCUAC GGAGGUGGCG 751 GCGGUAGACG CGAGGGUGGA GGAGGAUACA GCGGCGGCGG CGGCGGUUAC 801 UCCUCAAGAG GUGGUGGUGG CGGAAGCUAC GGUGGUGGAA GACGUGAGGG 851 AGGAGGAGGA UACGGUGGUG GUGAAGGAGG AGGUUACGGA GGAAGCGGUG 901 GUGGUGGAGG AUGGUAAUUC CUUUAAUUAG GUUUGGGAUU ACCAAUGAAU 951 GUUCUCUCUC UCGCUUGUUA UGCUUCUACU UGGUUUUGUG UGUUCUCUAU 1001 UUUGUUCUGG UUCUGCUUUA GAUUUGAUGU AACAGUUCGU GAUUAGGUAU 1051 UUUGGUAUCU GGAAACGUAA UGUUAAGUCA CUUGUCAUUC UCUAAAUAAC 1101 AAAUUUCUUC GGAGAUAUUA UCUCUGUUGA UUGAUUCUAU CAUCU

Sequence 2: AtGRP7 amino acid sequence.

Glycine stretch in gray, region used for structure prediction (cf. figure 2.16) in yellow with RNPs in red.

1 MASGDVEYRC FVGGLAWATD DRALETAFAQ YGDVIDSKII NDRETGRSRG 51 FGFVTFKDEK AMKDAIEGMN GQDLDGRSIT VNEAQSRGSG GGGGHRGGGG 101 GGYRSGGGGG YSGGGGSYGG GGGRREGGGG YSGGGGGYSS RGGGGGSYGG 151 GRREGGGGYG GGEGGGYGGS GGGGGW

(48)

(49)

3 Materials and Methods

3.1 Instrumentation

All instrumentation used in this work will be described in this chapter. This includes the different custom built spectroscopic setups used for confocal and widefield microscopy, as wells as standard commercial systems utilized.

3.1.1 Ensemble Spectroscopy Devices

3.1.1.1 Fluorescence Spectroscopy

A Cary Eclipse (Varian, Darmstadt, Germany) spectrometer was used for recording fluorescence spectra. The spectrometer is equipped with a thermoelectric peltier temperature control, allowing for thermal kinetic measurements.

Fluorescence spectra of recombinant Dronpa-s in solution

Purified Dronpa-s protein was diluted with PBS to 10-5M in a stan-dard quartz cuvette. The emission spectra of Dronpa were measured with the described fluorescence spectrometer at 488 nm excitation wavelength.

(50)

3.1.1.2 Absorption Spectroscopy

If light of the intensity I0 and the wavelength λ passes through a substance, it will be attenuated to the intensity I due to absorption. The attenuation depends on the concentration C of the substance, the specific absorption coefficient ε at the wavelength λ and d the length of the trajectory through the substance. If the concentration is below 10-6_{M, effects like fluorescence emission reabsorption and} intermolecular interactions can be neglected. Thus the absorption A linearly relates to the concentration In this case, the absorption of light is given by the Lambert-Beer law.

A= logI0

I = ε · C · d . (3.1)

Absorption spectra were obtained with a Lambda 25 absorption spec-trometer (PerkinElmer). The samples were diluted to a concentra-tion of 10-7_{M in quartz glass cuvettes (Hellma), prior to} examina-tion.

3.1.2 Confocal Laser Scanning Microscopes

In this work varying confocal systems were used for live cell imaging. Each has its own set of advantages and disadvantages. While the commercial setups provide out of the box functionality for standard experiments, it can be hard to get quantitative results or perform rather uncommon methods on them. This is where a custom-built setup comes into play, as it is highly scalable.

Confocal systems can be of two types in general. The most common used technique in commercial systems is the beam scanning approach, where the laser beam is deflected by galvanometer driven mirrors, thus scanning the static sample. Both commercial systems used in this work are of beam scanning type.

Another approach is realized via stage scanning. Here the whole stage harboring the sample is moved in relation to a static laser beam, also

(51)

3.1 Instrumentation

resulting in the scanning of the sample. Therefore, the stage has to be moved with high accuracy, which is often realized by a piezo driven stage. This technique was used for the custom-built setup, described in chapter 3.1.2.3.

3.1.2.1 Zeiss LSM 710

The LSM 710 (Carl Zeiss AG, Germany) is an inverse confocal micro-scope (see chapter 2.3), equipped with a Helium Neon (HeNe) Laser and an Argon Ionen (Ar+) Laser. Thus, it provides several excitation laser lines: Ar+ _{458 nm , 488 nm, and 514 nm. HeNe 543 nm and} 633 nm. Input laser powers of the Ar+ _{lines could only be adjusted} arbitrarily by setting a remote control knob. Input laser power of the HeNe could not be adjusted. Furthermore, the software provides a final laser power adjustment in relative units (0 – 100 %). Whereas 100 % corresponds to the input laser power. Due to the security mechanisms it is not possible to directly measure applied laser powers with an external device, as all laser emission was shut off when the optical path was interrupted by any means.

Input laser powers for each line were measured during the initial installation of the setup to be 1.42 mW for 458 nm, 7.18 mW for 488 nm, 5.76 mW for 514 nm, 0.56 mW for 543 nm and 1.71 mW for 633 nm. The Ar+ _{remote control knob was adjusted to the same maximum level at} its best. Therefore, laser power was assumed to be nearly the given values at initial setup attenuated by the respective percentage set in the software.

The LSM 710 facilitates two independent galvanometric scanning mirrors by which the laser beam is guided through a PlanAPO 63x/1.40 NA oil immersion objective (Zeiss), focusing on the sample. The fluorescence is collected by the same objective, transmits the dichroic filter and is further directed to a pinhole with adjustable aperture size. A subsequent grid spectrally separated the light nanometer-wise and directed it to three different photomultiplier tubes (PMT) . Prior to that, the spectral range could be narrowed to appropriate detection windows by multiple filter sets. The built-in z-drive (Axio Observer)

(52)

supplied 3D-scans of the samples with a minimum step size of below 25 nm.

The bundled evaluation software Zen (Zeiss) that came with the LSM 710 supplied a rich interface individual settings and a broad range of tools and algorithms for automated procedures, such as time-lapse imaging and also a FRAP analysis. For various reasons a manual postponed FRAP analysis was favored over the in-built tool (see chapter 3.2.2.2 on page 47).

Resulting images were saved as .lsm-files with a set of all applied measurement parameters for further processing. This file format can be imported by ImageJ (via plugin) and Fiji (see chapter 3.2.2 on page 46).

3.1.2.2 Leica TCM SP2

In general the TCM SP2 (Leica) features the same backbone com-ponents as the LSM 710 except for the upright orientation of the microscope. In addition to the laser lines mentioned in chapter 3.1.2.1 for the LSM, the SP2 features a 405 nm laser line. For imaging a PlanApo ×63; 1.1 NA oil immersion objective (Leica) was deployed. The according laser powers could be obtained by placing the sensor of an external power-meter (Field Master, Coherent) in the focal plane of the objective.

3.1.2.3 Custom Build CLSM

A custom-built confocal laser scanning microscope was used for fluores-cence imaging. The main components are schemed in figure 3.1 on the facing page. The setup consists of an inverted microscope (Axiovert 200M; Zeiss) equipped with a piezo stage (PI-509; Physik Instrumente). The piezo stage moves the sample through the focus and is controlled by an analog output card (PCI-6713; National Instruments). An argon-ion laser (Ion Laser Technology) emitting at 488 nm and a laser diode (Vioflame, Coherent) emitting at 405 nm are used for excitation. Both

(53)

3.1 Instrumentation Obj L LP Pinhole Ar+ APD APD APD APD TCSPC Card 405 nm M M S C BP Prism Telescope DCM Opt. Fiber Sample Piezo Stage x z y L L L L L BP BP BP DCM DCM DCM DCM Diaphragm

Figure 3.1: Scheme of the custom-built confocal laser scanning setup.

Optical elements used in the excitation path, indicated by the blue dashed box: Argon-Ion laser (Ar+_{) and a diode laser emitting at 405 nm as laser} sources. Collimator (C), mirror (M), lens (L), Kepler-type telescope and an optical fiber, were used for beam guiding and shaping. Both excitation lasers could be attenuated and shut independently (S) and were overlaid by a dichroic mirror (DCM). Prism and a diaphragm were used to separate the 488 nm line from the residual Ar+ laser lines.

Optical elemets within the detection path, indicated by the red dashed box: Long-pass (LP) and band-pass (BP) filters depleted raman and rayleigh scattered laser light. DCMs were used for spectral separation of the fluo-rescence signal. Avalanche photo diodes (APD) were used as detectors for the spectrally filtered light. For signal processing a time correlated single photon counting (TCSPC) PCI card was deployed.

(54)

laser sources are controlled by a shutter and a variable attenuator. The 488 nm laser light beam was extended using a telescope, overlaid with the 405 nm laser beam by the use of a dichroic mirror and coupled into the microscope where it was directed into the objective (Plan Apo, ×60, NA 1.4 oil; Olympus). The light beam was focused onto the sample and fluorescence light was collected by the same objective and spectrally separated using a dichroic beamsplitter (z 405/488; Chroma Technology). The parallel light beam was first focused onto a 100 µm pinhole and then focused onto the active area of an avalanche photodi-ode (AQR-16; Perkin Elmer) passing an emission filter (HQ 500 LP; Chroma Technology). The electric signal was then processed by a time-correlated single-photon counting device (SPC-630; Becker & Hickl). The whole set-up is controlled by a software based on LabView that controls the movement of the piezo scan stage and attributes photon information obtained by the photon-counting device to the appropriate x- and y -positions on the sample. If not stated otherwise, the piezo movement was adjusted to a dwell time of 1 ms/px and a step size of 500 nm/px for laser scanning experiments.

3.1.3 Fluorescence Correlation Spectroscopy Setup

FCS experiments were performed on a custom-made confocal fluores-cence microscope (see figure 3.2) that essentially consists of a standard inverse fluorescence microscope equipped with a HeNe laser, emitting at 632.8 nm, as excitation source and a pinhole of 100 µm size. The col-limated laser beam was coupled into an oil-immersion objective (×60; NA 1.35, Zeiss) by a dichroic beam splitter (645DLRP, Omega Optical, Brattleboro, VT). The average laser power was adjusted to be 0.5 mW before entering the aperture of the microscope, so as not to populate the triplet state of the fluorophores. The fluorescence signal was col-lected by the same objective, filtered by a band-pass filter (700RDF75, Omega Optical), and imaged onto the active area of two single-photon avalanche photodiodes (APDs) (SPCM-AQR-14, PerkinElmer Opto-electronics, Vaudreuil, QC, Canada), sharing the fluorescence signal by a cubic non-polarizing beamsplitter (Linos, Göttingen, Germany). The signals of the APDs were recorded at cross-correlation setting

(55)

3.1 Instrumentation

(9 cycles a 100 s for each measurement) by using a real-time photon correlator PCI card (DPC-230, Becker & Hickl, Berlin, Germany). The application of two APDs at cross-correlation setting circumvents dead-time and after-pulsing effects.

Obj L LP Pinhole APD APD _Correl. Card 633 nm M S C BP Opt. Fiber Sample L L L DCM Ex BS

Figure 3.2: Scheme of the custom-built FCS setup.

Optical elements used in the excitation path, indicated by the blue dashed box: Diode laser equipped with excitation filter (Ex) emitting at 633 nm as laser source. Collimator (C), mirrors (M), lenses (L), and an optical fiber, were used for beam guiding and shaping. An attenuator/shutter (S) was used for adjusting the laser intensity.

Optical elemets within the detection path, indicated by the red dashed box: Long-pass (LP) and band-pass (BP) filters depleted raman and rayleigh scattered laser light. A 50/50 beam splitter cube (BS) was used for equal distribution of the fluorescence signal on to APDs. For signal processing a correlator PCI card was deployed.

Advanced fluorescence based methods for protein interaction studies applied to the RNA binding protein AtGRP7

Advanced

Fluorescence based Methods

for Protein Interaction Studies

applied to the

RNA binding Protein AtGRP7

Dissertation

Inauguraldissertation

zur Erlangung der Doktorwürde

der Fakultät für Physik

der Universität Bielefeld

Advanced

Fluorescence based Methods

for Protein Interaction Studies

applied to the

RNA binding Protein AtGRP7

vorgelegt von

Fabian Humpert

geboren in Wickede

Zusammenfassung

Abstract

Contents

1 Introduction

1.1 Fluorescent proteins

1.2 Target Protein AtGRP7

1.3 Application of photoswitchable proteins

in standard imaging techniques

1.4 Colocalization analysis visualizing

protein interactions

2 Theoretical Background

2.1 Principles of Fluorescence

S

S

S

T

E

ne

rgy

E

ne

rgy

r

Intensity

Energy

0 6

0 5

0 4

0 2

0 3

0 1

0 0

0 6

5 0

4 0

2 0

3 0

1 0

6 0

Fluorescence

Absorption

2.2 Fluorescence Spectroscopy

2.2.1 Fluorophores

2.2.2 Fluorescence Quenching

Reduction

Oxidation

hν

Ener

gy

hν

2.2.3 Fluorescence Correlation Spectroscopy

G(τ)

log(τ)

2.3 Fluorescence Microscopy

2.4 Biological Background

2.4.1 Structure of Nucleic Acids

a

b

2.4.2 Protein Biosynthesis

Intron

5' UTR

_hν